The Morphosyntactic Wavelets, also known as MLW (Morphosyntactic Linguistic Wavelets) is a linguistic modeling tool that provides a dynamic description of any kind of text. Its name derives from the original concept of mathematical wavelets, a synthetic description of signals.
The MLW emerges as the result of a lack of general description of a language. The first approaches were based on specific language tips explicity implemented as dictionaries, tagging tools, specific languages and even complex structures that aimed to represent alternatively ontology or semantics. The lack of plasticity and complexity of these methods derived in a lack of popularity and even discrepancies in the core of the community. With the MLW, language is processed as a special “signal” that can be compressed, filtered and translated into other “axes” to take the same information from other perspectives.
Language requires steps that are similar to those required by traditional numeric wavelets. These steps are described in more detail in following wikis. In brief, these steps are:
Take the original text sample.
Compress and translate text into an oriented graph preserving most morphosyntactic properties
Apply filtering to get the better place within the current knowledge organization
If abstraction granularity and details are insufficient for the current information
Select a better filter, according to the knowledge organization
Repeat from step 3
Take the resulting sequence of filtering as a current representation of the knowledge about and ontology of the text
Take the resulting structure as the internal representation of the new status of knowledge
Morphosyntactic Linguistic Wavelets for Knowledge Management. Daniela López De Luise. InTech. Open Book. 2012.
Author
Daniela López De Luise